Nonconvex Quadratic Programming Method for k－Coloring Problem：Algorithm and Computation

In this paper, we consider the socalled k-coloring problem in general case.Firstly, a special quadratic 0-1 programming is constructed to formulate k-coloring problem. Secondly, by use of the equivalence between above quadratic0-1 programming and its relaxed problem, k-coloring problem is converted intoa class of (continuous) nonconvex quadratic programs, and several theoreticresults are also introduced. Thirdly, linear programming approximate algorithmis quoted and verified for this class of nonconvex quadratic programs. Finally,examining problems which are used to test the algorithm are constructed andsufficient computation experiments are reported.

最小权点覆盖问题的一个近似算法

在点、边赋权的简单图中,关于最小权点覆盖问题,以经典的最短路算法-Dijkstra算法为基础,提出了一个求解该问题的近似算法.首先,在给定的赋权图中任选一点作为初始点,并给出允许集及相关定义.然后,利用经典的最短路算法-Dijkstra算法,求出初始点到允许集中各顶点的最短路径,并按照一定的原则选择近似最小权点覆盖集.最后,通过算例阐释了算法的实现过程的合理性及有效性.